The complement system provides an early acting mechanism to initiate and amplify the inflammatory response to microbial infection and other acute insults (M. K. Liszewski and J. P. Atkinson, 1993, in Fundamental Immunology, Third Edition, edited by W. E. Paul, Raven Press, Ltd., New York). While complement activation provides a valuable first-line defense against potential pathogens, the activities of complement that promote a protective inflammatory response can also represent a potential threat to the host (K. R. Kalli, et al., Springer Semin. Immunopathol. 15:417-431, 1994; B. P. Morgan, Eur. J. Clinical Investig. 24:219-228, 1994). For example, C3 and C5 proteolytic products recruit and activate neutrophils. These activated cells are indiscriminate in their release of destructive enzymes and may cause organ damage. In addition, complement activation may cause the deposition of lytic complement components on nearby host cells as well as on microbial targets, resulting in host cell lysis.
The complement system has been implicated as contributing to the pathogenesis of numerous acute and chronic disease states, including: myocardial infarction, revascularization following stroke, ARDS, reperfusion injury, septic shock, capillary leakage following thermal burns, postcardiopulmonary bypass inflammation, transplant rejection, rheumatoid arthritis, multiple sclerosis, myasthenia gravis, and Alzheimer's disease. In almost all of these conditions, complement is not the cause but is one of several factors involved in pathogenesis. Nevertheless, complement activation may be a major pathological mechanism and represents an effective point for clinical control in many of these disease states. The growing recognition of the importance of complement-mediated tissue injury in a variety of disease states underscores the need for effective complement inhibitory drugs. No drugs have been approved for human use that specifically target and inhibit complement activation.
Currently, it is widely accepted that the complement system can be activated through three distinct pathways: the classical pathway, the lectin pathway, and the alternative pathway. The classical pathway is usually triggered by antibody bound to a foreign particle (i.e., an antigen) and thus requires prior exposure to that antigen for the generation of specific antibody. Since activation of the classical pathway is associated with development of an immune response, the classical pathway is part of the acquired immune system. In contrast, both the lectin and alternative pathways are independent of clonal immunity and are part of the innate immune system.
The first step in activation of the classical pathway is the binding of a specific recognition molecule, C1q, to antigen-bound IgG and IgM. The activation of the complement system results in the sequential activation of serine protease zymogens. C1q is associated with the C1r and C1s serine protease proenzymes as a complex called C1 and, upon binding of C1q to an immune complex, autoproteolytic cleavage of the Arg-I1e site of C1r is followed by C1r activation of C1s, which thereby acquires the ability to cleave C4 and C2. The cleavage of C4 into two fragments, designated C4a and C4b, allows the C4b fragments to form covalent bonds with adjacent hydroxyl or amino groups and the subsequent generation of C3 convertase (C4b2b) through noncovalent interaction with the C2b fragment of activated C2. C3 convertase (C4b2b) activates C3 leading to generation of the C5 convertase (C4b2b3b) and formation of the membrane attack complex (C5b-9) that can cause microbial lysis. The activated forms of C3 and C4 (C3b and C4b) are covalently deposited on the foreign target surfaces, which are recognized by complement receptors on multiple phagocytes.
Independently, the first step in activation of the complement system by the lectin pathway is also the binding of specific recognition molecules, which is followed by the activation of associated serine proteases. However, rather than the binding of immune complexes by C1q, the recognition molecules in the lectin pathway are carbohydrate-binding proteins (mannan-binding lectin (MBL), H-ficolin, M-ficolin, and L-ficolin) (J. Lu et al., Biochim. Biophys. Acta 1572:387-400, 2002; Holmskov et al., Annu. Rev. Immunol. 21:547-578 (2003); Teh et al., Immunology 101:225-232 (2000)). Ikeda et al. first demonstrated that, like C1q, MBL could activate the complement system upon binding to yeast mannan-coated erythrocytes in a C4-dependent manner (K. Ikeda et al., J. Biol. Chem. 262:7451-7454, 1987). MBL, a member of the collectin protein family, is a calcium-dependent lectin that binds carbohydrates with 3- and 4-hydroxy groups oriented in the equatorial plane of the pyranose ring. Prominent ligands for MBL are thus D-mannose and N-acetyl-D-glucosamine, while carbohydrates not fitting this steric requirement have undetectable affinity for MBL (Weis, W. I., et al., Nature 360:127-134, 1992). The interaction between MBL and monovalent sugars is extremely weak, with dissociation constants typically in the 2 mM range. MBL achieves tight, specific binding to glycan ligands by interaction with multiple monosaccharide residues simultaneously (Lee, R. T., et al., Archiv. Biochem. Biophys. 299:129-136, 1992). MBL recognizes the carbohydrate patterns that commonly decorate microorganisms such as bacteria, yeast, parasites and certain viruses. In contrast, MBL does not recognize D-galactose and sialic acid, the penultimate and ultimate sugars that usually decorate “mature” complex glycoconjugates present on mammalian plasma and cell surface glycoproteins. This binding specificity is thought to help protect from self activation. However, MBL does bind with high affinity to clusters of high-mannose “precursor” glycans on N-linked glycoproteins and glycolipids sequestered in the endoplasmic reticulum and Golgi of mammalian cells (Maynard, Y., et al., J. Biol. Chem. 257:3788-3794, 1982). Therefore, damaged cells are potential targets for lectin pathway activation via MBL binding.
The ficolins possess a different type of lectin domain than MBL, called the fibrinogen-like domain. Ficolins bind sugar residues in a Ca++-independent manner. In humans, three kinds of ficolins, L-ficolin, M-ficolin and H-ficolin, have been identified. Both serum ficolins L-ficolin and H-ficolin have in common a specificity for N-acetyl-D-glucosamine; however, H-ficolin also binds N-acetyl-D-galactosamine. The difference in sugar specificity of L-ficolin, H-ficolin and MBL means that the different lectins may be complementary and target different, though overlapping, glycoconjugates. This concept is supported by the recent report that, of the known lectins in the lectin pathway, only L-ficolin binds specifically to lipoteichoic acid, a cell wall glycoconjugate found on all Gram-positive bacteria (Lynch, N. J., et al., J. Immunol. 172:1198-1202, 2004). The collectins (i.e., MBL) and the ficolins bear no significant similarity in amino acid sequence. However, the two groups of proteins have similar domain organizations and, like C1q, assemble into oligomeric structures, which maximize the possibility of multisite binding. The serum concentrations of MBL are highly variable in healthy populations and this is genetically controlled by the polymorphism/mutations in both the promoter and coding regions of the MBL gene. As an acute phase protein, the expression of MBL is further upregulated during inflammation. L-ficolin is present in serum at similar concentrations as MBL. Therefore, the L-ficolin arm of the lectin pathway is potentially comparable to the MBL arm in strength. MBL and ficolins can also function as opsonins, which require interaction of these proteins with phagocyte receptors (Kuhlman, M., et al., J. Exp. Med. 169:1733, 1989; Matsushita, M., et al., J. Biol. Chem. 271:2448-54, 1996). However, the identities of the receptor(s) on phagocytic cells have not been established.
Human MBL forms a specific and high affinity interaction through its collagen-like domain with unique C1r/C1s-like serine proteases, termed MBL-associated serine proteases (MASPs). To date, three MASPs have been described. First, a single enzyme “MASP” was identified and characterized as the enzyme responsible for the initiation of the complement cascade (i.e., cleaving C2 and C4) (Ji, Y. H., et al., J. Immunol. 150:571-578, 1993). Later, it turned out that MASP is in fact a mixture of two proteases: MASP-1 and MASP-2 (Thiel, S., et al., Nature 386:506-510, 1997). However, it was demonstrated that the MBL-MASP-2 complex alone is sufficient for complement activation (Vorup-Jensen, T., et al., J. Immunol. 165:2093-2100, 2000). Furthermore, only MASP-2 cleaved C2 and C4 at high rates (Ambrus, G., et al., J. Immunol. 170:1374-1382, 2003). Therefore, MASP-2 is the protease responsible for activating C4 and C2 to generate the C3 convertase, C4b2b. This is a significant difference from the C1 complex, where the coordinated action of two specific serine proteases (C1r and C1s) leads to the activation of the complement system. Recently, a third novel protease, MASP-3, has been isolated (Dahl, M. R., et al., Immunity 15:127-35, 2001). MASP-1 and MASP-3 are alternatively spliced products of the same gene. The biological functions of MASP-1 and MASP-3 remain to be resolved.
MASPs share identical domain organizations with those of C1r and C1s, the enzymatic components of the C1 complex (Sim, R. B., et al., Biochem. Soc. Trans. 28:545, 2000). These domains include an N-terminal C1r/C1s/sea urchin Vegf/bone morphogenic protein (CUB) domain, an epidermal growth factor-like domain, a second CUB domain, a tandem of complement control protein domains, and a serine protease domain. As in the C1 proteases, activation of MASP-2 occurs through cleavage of an Arg-I1e bond adjacent to the serine protease domain, which splits the enzyme into disulfide-linked A and B chains, the latter consisting of the serine protease domain. Recently, a genetically determined deficiency of MASP-2 was described (Stengaard-Pedersen, K., et al., New Eng. J. Med. 349:554-560, 2003). The mutation of a single nucleotide leads to an Asp-Gly exchange in the CUB1 domain and renders MASP-2 incapable of binding to MBL.
MBL is also associated with a nonenzymatic protein referred to as MBL-associated protein of 19 kDa (MAp19) (Stover, C. M., J. Immunol 162:3481-90, 1999) or small MBL-associated protein (sMAP) (Takahashi, M., et al., Int. Immunol. 11:859-863, 1999). MAp19 is formed by alternative splicing of the MASP 2 gene product and comprises the first two domains of MASP-2, followed by an extra sequence of four unique amino acids. The MASP 1 and MASP 2 genes are located on chromosomes 3 and 1, respectively (Schwaeble, W., et al., Immunobiology 205:455-466, 2002).
Several lines of evidence suggest that there are different MBL-MASPs complexes and a large fraction of the total MASPs in serum is not complexed with MBL (Thiel, S., et al., J. Immunol. 165:878-887, 2000). Both H- and L-ficolin are associated with MASP and activate the lectin complement pathway, as does MBL (Dahl, M. R., et al., Immunity 15:127-35, 2001; Matsushita, M., et al., J. Immunol. 168:3502-3506, 2002). Both the lectin and classical pathways form a common C3 convertase (C4b2b) and the two pathways converge at this step.
The lectin pathway is widely thought to have a major role in host defense against infection. Strong evidence for the involvement of MBL in host defense comes from analysis of patients with decreased serum levels of functional MBL (Kilpatrick, D. C., Biochim. Biophys. Acta 1572:401-413, 2002). Such patients display susceptibility to recurrent bacterial and fungal infections. These symptoms are usually evident early in life, during an apparent window of vulnerability as maternally derived antibody titer wanes, but before a full repertoire of antibody responses develops. This syndrome often results from mutations at several sites in the collagenous portion of MBL, which interfere with proper formation of MBL oligomers. However, since MBL can function as an opsonin independent of complement, it is not known to what extent the increased susceptibility to infection is due to impaired complement activation.
Although there is extensive evidence implicating both the classical and alternative complement pathways in the pathogenesis of non-infectious human diseases, the role of the lectin pathway is just beginning to be evaluated. Recent studies provide evidence that activation of the lectin pathway can be responsible for complement activation and related inflammation in ischemia/reperfusion injury. Collard et al. (2000) reported that cultured endothelial cells subjected to oxidative stress bind MBL and show deposition of C3 upon exposure to human serum (Collard, C. D., et al., Am. J. Pathol. 156:1549-1556, 2000). In addition, treatment of human sera with blocking anti-MBL monoclonal antibodies inhibited MBL binding and complement activation. These findings were extended to a rat model of myocardial ischemia-reperfusion in which rats treated with a blocking antibody directed against rat MBL showed significantly less myocardial damage upon occlusion of a coronary artery than rats treated with a control antibody (Jordan, J. E., et al., Circulation 104:1413-1418, 2001). The molecular mechanism of MBL binding to the vascular endothelium after oxidative stress is unclear; a recent study suggests that activation of the lectin pathway after oxidative stress may be mediated by MBL binding to vascular endothelial cytokeratins, and not to glycoconjugates (Collard, C. D., et al., Am. J. Pathol. 159:1045-1054, 2001). Other studies have implicated the classical and alternative pathways in the pathogenesis of ischemia/reperfusion injury and the role of the lectin pathway in this disease remains controversial (Riedermann, N. C., et al., Am. J. Pathol. 162:363-367, 2003).
In contrast to the classical and lectin pathways, no initiators of the alternative pathway have been found to fulfill the recognition functions that C1q and lectins perform in the other two pathways. Currently it is widely accepted that the alternative pathway is spontaneously triggered by foreign or other abnormal surfaces (bacteria, yeast, virally infected cells, or damaged tissue). There are four plasma proteins directly involved in the alternative pathway: C3, factors B and D, and properdin. Proteolytic generation of C3b from native C3 is required for the alternative pathway to function. Since the alternative pathway C3 convertase (C3bBb) contains C3b as an essential subunit, the question regarding the origin of the first C3b via the alternative pathway has presented a puzzling problem and has stimulated considerable research.
C3 belongs to a family of proteins (along with C4 and α-2 macroglobulin) that contain a rare posttranslational modification known as a thioester bond. The thioester group is composed of a glutamine whose terminal carbonyl group is bound to the sulfhydryl group of a cysteine three amino acids away. This bond is unstable and the electrophilic carbonyl group of glutamine can form a covalent bond with other molecules via hydroxyl or amino groups. The thioester bond is reasonably stable when sequestered within a hydrophobic pocket of intact C3. However, proteolytic cleavage of C3 to C3a and C3b results in exposure of the highly reactive thioester bond on C3b and by this mechanism C3b covalently attaches to a target. In addition to its well-documented role in covalent attachment of C3b to complement targets, the C3 thioester is also thought to have a pivotal role in triggering the alternative pathway. According to the widely accepted “tick-over theory”, the alternative pathway is initiated by the generation of a fluid-phase convertase, iC3Bb, which is formed from C3 with hydrolyzed thioester (iC3; C3(H2O)) and factor B (Lachmann, P. J., et al., Springer Semin. Immunopathol. 7:143-162, 1984). The C3b-like iC3 is generated from native C3 by a slow spontaneous hydrolysis of the internal thioester in the protein (Pangburn, M. K., et al., J. Exp. Med. 154:856-867, 1981). Through the activity of the iC3Bb convertase, C3b molecules are deposited on the target surface thereby initiating the alternative pathway.
Very little is known about the initiators of activation of the alternative pathway. Activators are thought to include yeast cell walls (zymosan), many pure polysaccharides, rabbit erythrocytes, certain immunoglobulins, viruses, fungi, bacteria, animal tumor cells, parasites, and damaged cells. The only feature common to these activators is the presence of carbohydrate, but the complexity and variety of carbohydrate structures has made it difficult to establish the shared molecular determinants, which are recognized.
The alternative pathway can also provide a powerful amplification loop for the lectin/classical pathway C3 convertase (C4b2b) since any C3b generated can participate with factor B in forming additional alternative pathway C3 convertase (C3bBb). The alternative pathway C3 convertase is stabilized by the binding of properdin. Properdin extends the alternative pathway C3 convertase half-life six to ten fold. Addition of C3b to the C3 convertase leads to the formation of the alternative pathway C5 convertase.
All three pathways (i.e., the classical, lectin and alternative) have been thought to converge at C5, which is cleaved to form products with multiple proinflammatory effects. The converged pathway has been referred to as the terminal complement pathway. C5a is the most potent anaphylatoxin, inducing alterations in smooth muscle and vascular tone, as well as vascular permeability. It is also a powerful chemotaxin and activator of both neutrophils and monocytes. C5a-mediated cellular activation can significantly amplify inflammatory responses by inducing the release of multiple additional inflammatory mediators, including cytokines, hydrolytic enzymes, arachidonic acid metabolites and reactive oxygen species. C5 cleavage leads to the formation of C5b-9, also known as the membrane attack complex (MAC). There is now strong evidence that sublytic MAC deposition may play an important role in inflammation in addition to its role as a lytic pore-forming complex.